Mbps vs MB/s: convert internet speeds correctly

Internet access rates are usually advertised in megabits per second, written Mbps or Mb/s. File sizes and download programs often use megabytes and MB/s. A byte contains eight bits, so the numbers differ by a factor of about eight even before protocol overhead is considered.

Bits, bytes and a quick conversion

Lowercase b denotes a bit, while uppercase B denotes a byte. With decimal units, 1 Mbps is 1,000,000 bits per second and 1 MB/s is 1,000,000 bytes per second. A 100 Mbps connection therefore has a theoretical data rate of 12.5 MB/s: divide megabits by eight. To go the other way, multiply MB/s by eight. Capitalisation matters; Mbps and MB/s are not interchangeable.

File and memory sizes may also use binary units. One MiB equals 1,048,576 bytes. Some software labels a binary value “MB”, making a smaller additional difference when results are compared. Check both the symbol and the counting convention, especially when the disagreement is only a few per cent.

Estimate a download

A 1 GB file in decimal units contains about 8 gigabits. At a sustained useful rate of 100 Mbps, the ideal transfer time is about 80 seconds. In reality, Ethernet, IP, TCP or QUIC, encryption and the application add overhead. The remote server may send more slowly as well. An advertised plan usually describes a maximum or stated line rate under particular conditions, not a promise for every file.

If a browser reports 10 MB/s instead of an expected “100”, convert the units first: 10 MB/s is about 80 Mbps. Then investigate Wi-Fi, other devices using the connection, server limits and upload versus download direction. A ping time in milliseconds cannot be compared with a transfer rate in bytes per second; they measure different properties.

Measure under useful conditions

Run the 2ip speed test several times, preferably over Ethernet with other large transfers paused. Check download and upload separately; they may differ greatly. For a file transfer, use a sufficiently long period so startup time and caches do not dominate the average. Compare results with the plan using the same units.

For a call or game, throughput alone is insufficient. Consider delay, variation in delay and packet loss as explained in the latency and jitter guide. Always write units next to a number so a notation difference is not mistaken for a network fault.

NIST distinguishes decimal prefixes such as mega from binary ones such as mebi. This is why MB and MiB differ even when describing the same number of bytes.